Pattern recognition of Nebbiolo wine as a function of the fungicide treatment of the grape

Principal component analysis and classification techniques were used to inv estigate the effect on the properties of wine of different fungicide treatm ents performed on the grapes. 55 variables were determined, 17 on the grape just-collected and the others on the wines. 27 samples of grape and wine w ere considered, produced in 1992, 1993 and 1994. The wines were classified as a function of the production year but no important correlation was found between fungicide treatment and the wine features. The classification was performed by a SIMCA method. Clustering analysis, on the other hand, allowed a separation of the wine sa mples in three clusters, relative to three main kinds of grape treatments, i.e. based on i) copper sulfate, ii) copper hydroxide and iii) Mancozeb (th e only one that does not contain Cu). The treatments based on copper oxychl oride and the blank; that are the less active or characterized treatments, gave mixed results within the three clusters.